See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/222407096

Measurement and analysis of rock mass discontinuity spacing and frequency

in part of the Fimiston Open Pit operation in Kalgoorlie, Western Australia: A
case study

Article  in  International Journal of Rock Mechanics and Mining Sciences · July 2002

DOI: 10.1016/S1365-1609(02)00003-5

CITATIONS READS

36 814

2 authors, including:

David Wines
Itasca Australia Pty Ltd
8 PUBLICATIONS   77 CITATIONS   

SEE PROFILE

All content following this page was uploaded by David Wines on 26 February 2019.

The user has requested enhancement of the downloaded file.

 International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602

Measurement and analysis of rock mass discontinuity spacing and

frequency in part of the Fimiston Open Pit operation in Kalgoorlie,
Western Australia: a case study
D.R. Winesa, P.A. Lillyb,*
a
Sinclair Knight Merz Pty Ltd, Melbourne, Australia
b
Western Australian School of Mines, Curtin University of Technology, PMB 22 Kalgoorlie, WA 6430, Australia
Accepted 19 December 2001

Abstract

The Fimiston Open Pit is located in the Eastern Goldfields of Western Australia and is mining an area traditionally known as the
Golden Mile. The geological succession in part consists of a 3000–4000 m thick sequence of ultramafic to mafic volcanic rocks and
mafic sills, including the Golden Mile Dolerite and the Paringa Basalt into which 650 m high pit slopes will ultimately be excavated.
A total of 349 m of scanline were mapped (770 discontinuity measurements), 1022 m of core logged (a further 1795 discontinuity
measurements) and a detailed analysis of discontinuity spacing and frequency were undertaken.
The total discontinuity spacing values for the study area clearly follow the shape of the exponential distribution, which is
consistent with many previous workers’ results. Although not as clearly defined, the normal set spacing values also appear to follow
the exponential distribution. The total average calculation for rock quality designation (RQD*) in this study is within 5% of the
total average estimate for RQD. This confirms that the method proposed by Priest and Hudson (Int J Rock Mech Min Sci &
Geomech Abstr 13 (1976) 135) provides a valuable tool that would be particularly beneficial for geotechnical investigations where
diamond core is not readily available.
Discontinuity spacing results are complicated by the influence of quartz veining and foliation planes. The veins appear to have
influenced the scanline mapping data to a greater extent than the core logging data, while the foliation appears to have influenced the
core logging data to a greater degree. r 2002 Elsevier Science Ltd. All rights reserved.

1. Introduction 1989 with the formation of Kalgoorlie Consolidated

Gold Mines Pty Ltd. (KCGM), a joint venture between
1.1. Background Normandy Mining Ltd. and Homestake Gold of
Australia Ltd.
The city of Kalgoorlie-Boulder is situated B600 km
to the east of Perth in Western Australia. The Fimiston
Open Pit (or ‘Super Pit’) is located on the eastern 1.2. Geology
boundary of the city and is mining an area traditionally
known as the Golden Mile. The area has been Clout and others [1] provide a comprehensive
continuously worked since 1893, predominantly by description of the geology in this part of Western
underground mining methods and more recently Australia and much of the information provided here
through open cut development. Various small-scale has been sourced from this work. The Kalgoorlie
open pits were developed in the early to mid-1980s goldfields are situated within the north–northwest
and the amalgamation of these pits was realised in May striking Archaean Norseman-Wiluna Greenstone Belt.
The succession consists of a 3000–4000 m thick sequence
*Corresponding author. Tel.: +61-8-9088-6150; fax: +61-8-9088- of ultramafic to mafic volcanic rocks and mafic sills,
6151. which is overlain by a 1 km thick sequence of volcano-
E-mail address: lillyp@wasm.curtin.edu.au (P.A. Lilly). sedimentary rocks. The gold deposits are predominantly

1365-1609/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 1 3 6 5 - 1 6 0 9 ( 0 2 ) 0 0 0 0 3 - 5
590 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602

hosted by the ultramafic and mafic rocks. The main rock One of the major stages in the development of the east
types encountered in the area are as follows: wall of the pit is the Oroya cutback (Fig. 1). The wall
The Golden Mile Dolerite (GMD) is a differentiated design involves either 20 or 30 m high batter faces, with
gabbroic sill, with a thickness ranging from 600 to the former being used in areas where earlier under-
940 m. The sill is divided into 10 sub units depending ground workings are encountered. All batters are
upon composition. The GMD is the most important designed at 651 and are separated by 10 m wide catch
gold host in the area. berms. PB, GMD and PD rock masses will be
encountered during the development of this stage and
* The Paringa Basalt (PB) consists of pillowed to wall development will occur in fresh material only. It is
massive basalts with minor thin interflow sediments. the analysis of defect spacing and frequency within these
The unit is B850 m thick. rock masses that is the subject of this paper.
* The Black Flag Beds (BFB) is a sequence of
sediments and felsic volcanics, with a thickness of
up to 1000 m. The thickness is reduced in the vicinity 2. Field data collection
of the Golden Mile due to the influence of folding.
* Porphyry Dykes (PD) crosscut all formations in the 2.1. Scanline mapping
stratigraphic sequence and range from 0.5 to 10 m in
width in the study area. Ten linear scanlines were undertaken, the details and
locations of which are presented in Table 1 and Fig. 1,
There is some debate as to the structural history of the respectively. A total of 349 m of scanline were mapped
area. However, the following deformation events have for a total of 770 discontinuity measurements and the
been postulated [1]: following details recorded:

* D1 deformation resulted in tight northwest-striking, * Rock type.

upright folds and major sub-vertical northwest- * The intersection distance of each discontinuity with
striking faults, sub parallel to the lithological the scanline.
boundaries. Displacement along the faults is in the * Discontinuity orientations (dip and dip direction).
order of several kilometres. * Visible trace length and the type of termination.
* The faulted and folded sequence created by the D1 * RMR parameter (‘‘Condition of Joints’’).
deformation was later crosscut by north-striking, * Q parameters (‘‘Jn’’, ‘‘Jr’’ and ‘‘Ja’’).
steeply west-dipping faults of the D2 phase. Displace- * GSI parameters (‘‘Structure’’ and ‘‘Surface Condi-
ment along the faults varies from several hundred tions’’).
metres to 2 kilometres. The mineralisation in the * Joint roughness coefficient (JRC).
study area was emplaced into the shear zones that * Joint compressive strength (JCS).
were active during this deformation phase. * Infill information.
* D3 deformation resulted in an upright foliation being
formed. Scanline mapping provided a significant amount of
valuable geotechnical information and was particularly
Quartz veining is also found throughout the rock
well suited to the study area where significant lengths of
mass in the study area and appears to post-date the
wall exposure are available. The orientations of the
mineralisation.
scanlines relative to the orientations of the major joint
sets (Fig. 2) ensured that sampling bias for three of the
1.3. Mining four sets was minimised. However, the following draw-
backs in the method are noted in the context of this
The pit is being mined in a series of cutbacks and, study:
based on current plans, the depth of the ultimate pit will
exceed 650 m. Pre-splitting is currently used along all * The exposures available for mapping were relatively
final and interim batter faces in the transitional and remote from the proposed cutback location.
fresh material. Modified production blasting is used * Not all exposed faces were accessible due to unsafe
adjacent to the design pit limits and involves the use of berm and floor conditions.
angled and smaller diameter blast holes in the one or * With a high percentage of the available wall
two rows closest to the design batter face. Three exposures striking north south, the ability to run
Komatsu PC8000 face shovels undertake most of the orthogonal scanlines was limited. Also, a near
earthmoving and an operating bench height of 10 m is horizontal joint set (Fig. 2) is found throughout the
generally used. A backhoe-configured excavator pit. However, the near-vertical scanlines necessary to
(EX1100) is used to scale the batter faces. minimise the bias for this set are impractical.
 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602 591

Fig. 1. Cutback design and location of scanline mapping traverses (with ON, OY or OP prefixes) and diamond drill holes (all other prefixes) for
which geotechnical logs are available.

* Although pre-splitting is used along most walls, each * Quartz veins occur throughout the rock mass in the
batter face exhibits some degree of blast damage. Oroya cutback area and deciding which veins to
Damage will often lead to induced cracking and care record as discontinuities proved to be difficult. It
had to be taken not to record such cracks as might be possible to inspect diamond core from the
discontinuities. It was generally found that the area and note the characteristics of those veins that
induced cracks could be identified and therefore cause breakage of the core. These characteristics
ignored. They are of random orientation and are less could then be used as a decision making criterion
continuous and less weathered than the genuine when face mapping. However, the core available for
discontinuities. this study was obtained from holes that were drilled
592 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602

Table 1
Scanline mapping details for the Oroya cutback investigation

Scanline ID Rock type Plunge (1) Azimuth (1) Mapped length (m)

ON130T1 PB 0 75 13
ON130T2 PB 0 180 38
OY140T1 GMD 0 175 41
OY140T2 GMD 0 186 50
OY150T1 GMD 0 239 47
OP230T1 GMD 0 265 46
OP240T1 GMD 0 204 22
OP240T2 GMD 0 40 26
OP240T3 GMD 0 184 37
OP240T4 GMD 0 3 29

Fig. 2. Orientations of major defect sets, drill holes and mapping traverses.

near parallel to the predominant quartz vein orienta- proved to be difficult. Again, a better understanding
tions and therefore intersection with the veins is rare. of the influence of the foliation on the rock mass may
Also, core has since been split for assay sampling and be gained from the inspection of fresh diamond core.
breakage would have been induced during the
splitting process. 2.2. Diamond core logging
* Foliation occurs throughout the rock mass in the
Oroya area and the decision as to which foliation The details of the diamond holes used in this study are
planes should be recorded as discontinuities also presented in Table 2 and Fig. 1. The orientations of the
 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602 593

Table 2
Diamond drill hole details for the Oroya cutback investigation

Hole no. Date drilled Plunge (1) Azimuth (1) Core size Length (m) Information derived

HVGD002 May-97 59 095 HQ3/NQ2 180 Geotechnical log

HVGD003 Apr-95 60 090 NQ2 200 Geotechnical log
ORGD001 May-95 61 089 HQ3 150 Geotechnical log
HMGD009 May-95 70 090 NQ2 149 Geotechnical log
HMGD010 May-95 60 087 NQ2 236 Geotechnical log
HMGD011 Jun-95 60 089 HQ3/NQ2 235 Geotechnical log
PAGD010 Aug-95 59 094 HQ3 250 Geotechnical log
GTD01 Jan-91 50 049 HQ3 181 Lab samples only
GTD02 Jan-91 49 040 HQ3 181 Lab samples only
GTD04 Feb-91 55 267 HQ3 250 Lab samples only
GTD05 Feb-91 49 052 HQ3 250 Lab samples only
GTD07 Mar-91 60 057 F 293 Lab samples only
GTD08 Apr-91 51 052 HQ3 169 Lab samples only
GTD17 Nov-91 50 232 F 83 Lab samples only
GTD18 Nov-91 50 232 F 220 Lab samples only
GTD20 Dec-91 50 048 HQ3 295 Lab samples only
GTD26 Sep-92 50 090 HQ3 250 Lab samples only
GTD27 Oct-92 55 090 HQ3 179 Lab samples only
GTD36 May-93 49 091 F 152 Lab samples only
GTD37 May-93 54 089 F 230 Lab samples only
GTD47 Mar-94 50 270 F 195 Orientation data only
GTD48 Apr-94 50 270 F 152.00 Orientation data only
GTD49 Apr-94 55 270 F 155 Orientation data only
GTD54 F 47 101 HQ3 263 Lab samples only
GTD55 Oct-95 60 103 HQ3 145 Log and lab samples
GTD56 Oct-95 50 104 HQ3 199 Log and lab samples
OTGD015 Dec-96 62 136 HQ3 146 Lab samples only
OTGD016 Dec-96 62 125 HQ3 298 Lab samples only
HRD59 F 53 089 HQ3 405 Lab samples only

logged holes relative to the major joint sets (Fig. 2) are 3. Discontinuity spacing
such that structures were well sampled, with the
exception of the steeply dipping, east–west striking set.
3.1. Introduction
A total of 1022 m of core were logged with 1795
discontinuity measurements taken. Similar information
Discontinuity spacing is a basic measurement of the
to that gleaned from scanline mapping was obtained
distance between one discontinuity and another. The
from diamond core logging, except for the trace length
scope of discontinuity spacing determination is to
details. It is important to note that only those natural
ascertain the size of the blocks that make up the rock
features that caused separation of the core were
mass [2] because wall stability is strongly influenced by
recorded as discontinuities.
this block size. The discontinuity spacing parameter is
Diamond drill holes often allow information to be
often used in classification schemes such as the rock
gained from the immediate vicinity of the proposed
mass rating system.
slope design. Unfortunately, this was not particularly
Discontinuity spacing measurements can be separated
true for the study area with all of the holes having been
into the following three forms [3]:
planned to investigate an earlier design proposal, leaving
the logged holes a relatively large distance (up to 300 m) * Total spacing is the distance between two adjacent
from the current cutback design. However, most holes discontinuities, measured along a line of general, but
are oriented at relatively high angles to the dominant specified, location and orientation.
structures, which helped to minimise bias. * Set spacing is the spacing between two adjacent
Although the effects of blast damage are eliminated, discontinuities from a particular discontinuity set,
care must be taken to identify artificial cracks that have measured along a line of general, but specified,
been introduced during and after the drilling process. location and orientation.
This was found to be relatively simple, with the artificial * Normal set spacing is the set spacing measured along
cracks generally exhibiting greater roughness and less a line that is normal to the mean orientation of a
weathering than the natural discontinuities. particular set.
594 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602

The first two parameters can be directly estimated than the length of the sampling line may not be
from a linear discontinuity survey such as a scanline. encountered by that line.
The normal set spacing Xn can be determined from [3] * Imprecision is caused by the tendency of small
samples to produce mean values that exhibit incon-
Xn ¼ Xd cos d; ð1Þ
sistent random deviations from the true population
where Xd is a random value of set spacing for a set mean. The precision of mean discontinuity spacing
whose normal makes an acute angle d with a linear estimates will improve with an increase in the number
survey. of measured and calculated discontinuity spacings.

3.2. General aspects of discontinuity spacing distribution

Priest and Hudson [10] proposed a method for
estimating the precision of mean discontinuity spacing
For a linear survey the discontinuity intersection
estimates. If scanline mapping is undertaken in the same
points may be evenly spaced, clustered, random or a
direction at several different locations within a given
combination of these. Priest and Hudson [4] studied the
rock mass domain then, based on the central limit
results of comprehensive scanline mapping from three
theorem, the frequency distribution of the resulting
different tunnels in the UK. The rock types encountered
mean spacing values will tend to follow a normal
were chalk, limestone, sandstone and mudstone and it
distribution irrespective of the distribution of disconti-
was found that, unless there is a large predominance of
nuity spacings in the population. For a normal
evenly spaced discontinuities, any combination of evenly
distribution, some proportion fðzÞ of the various
spaced, clustered or randomly positioned discontinuities
mapping lines will yield a mean value within the range:
will produce an exponential distribution of spacing. The pffiffiffi
exponential distribution may be expressed as c7z SD= n; ð3Þ
f ðxÞ ¼ lelx ; ð2Þ where C is the true population mean, SD is the standard
where f ðxÞ is the frequency of a discontinuity spacing x; deviation, n is the number of discontinuity spacing
and l is the average number of discontinuities per metre. values and z is the standard normal variable associated
The mean and standard deviation (SD) are both equal to with a particular confidence level.
1=l: That is, for discontinuity data measured along a
Priest and Hudson’s findings [4] have been supported scanline, the mean value has a fðzÞ probability (or
by others; for example, [5–7]. However, Sen and Kazi [8] percentage probability given by 100fðzÞ) of being
recommend the use of a log normal distribution for the situated within the range provided by Eq. (3). Selected
analysis of discontinuity spacing estimates and reported values of z and fðzÞ are provided in Table 3.
that it provides greater flexibility because both the Where discontinuity spacings follow an exponential
average discontinuity spacing and the variance of the distribution, the formula can be reduced to [10]
discontinuity spacings are considered. n ¼ ðz=AÞ2 ; ð4Þ
It should be noted that the aforementioned work has
concentrated on total discontinuity spacing values. where A is the allowable proportionate error. This
Priest [3] reported that it is yet to be shown whether equation can be used to determine the sample size
the distributions of set spacings or normal set spacings required to achieve a given error bandwidth.
are exponential. Hudson and Priest [9] demonstrated Priest and Hudson [10] refer only to scanlines in the
that the exponential distribution of total spacing can be reviewed discussions. However, diamond core logging
obtained from the superimposition of non-exponential also provides a linear discontinuity survey and this study
distributions of set spacing. assumes that the methods for estimating the precision of
mean spacing estimates are also applicable to borehole
3.3. General considerations of inaccuracy and imprecision
in mean spacing estimates
Table 3
Values of fðzÞ for the normal distribution [10]
Priest [3] described two separate criteria by which the
reliability of mean discontinuity spacing estimates can z fðzÞ
be assessed: 0.675 0.50
0.842 0.60
* Inaccuracy is caused by the tendency of the mean 1.036 0.70
estimate to be biased by some persistent factor that 1.282 0.80
causes the result to be consistently in error. Short 1.645 0.90
1.960 0.95
sampling lines may produce inaccuracy where dis-
2.576 0.99
continuity sets with average spacing values greater
 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602 595

data. Priest and Hudson [4] suggested that the scanline and OP240T4) is presented in Fig. 4. Each data set was
length should be at least 50 times the mean discontinuity obtained from the combination of total spacing values
spacing, while Call and others [5] stated that measure- from mapping lines of similar orientation. Frequency
ment of at least 100 points per discontinuity set provides histograms for the normal set spacing values of joint sets
an ideal representation of the properties of that set. 1, 2 and 4 are presented in Figs. 5–7, respectively.

3.4. Discontinuity spacing data in the study area 3.5. Comparison between core logging and scanline
mapping data
Total spacing estimates determined from scanline
mapping and core logging are presented in Tables 4 and The diamond holes were drilled across the strike of
5, respectively. The similarity between mean and SD the predominant structural orientations whereas eight
values, particularly in Table 5, is consistent with the out of the 10 scanlines were mapped approximately
characteristics of the exponential distribution (Eq. (2)). parallel to the predominant geological structures. It
The set spacing and normal set spacing values for the would therefore be reasonable to expect the diamond
four main joint sets in the study area are presented in holes to intersect the predominant orientations more
Table 6. The latter values were estimated using Eq. (1). regularly than the scanlines, resulting in a lower average
A frequency plot for the total discontinuity spacing and maximum total discontinuity spacing for the core
values estimated from core logging is presented in Fig. 3 logging data.
while a similar plot for the total spacing data obtained However, the data in Tables 4 and 5 show that
from the north–south striking scanlines (ON130T2, average total discontinuity spacing estimates produced
OY140T1, OY140T2, OP240T1, OP240T2, OP240T3 by both core logging and scanline mapping is 0.45 m,

Table 4
Total spacing values for the scanline mapping data

Scanline no. Rock type Mapped length (m) Total joint spacing (m)

Mean SD Min Max

ON130T1 PB 13.00 0.24 0.23 o0.01 1.00

ON130T2 PB 38.00 0.53 0.44 o0.01 1.60
OY140T1 GMD 41.00 0.56 0.58 o0.01 3.30
OY140T2 GMD 50.00 0.76 0.59 o0.01 3.00
OY150T1 GMD 47.00 0.36 0.34 o0.01 2.30
OP230T1 GMD 46.00 0.35 0.33 o0.01 1.70
OP240T1 GMD 22.00 0.27 0.19 o0.01 0.80
OP240T2 GMD 26.00 0.31 0.27 o0.01 1.20
OP240T3 GMD 37.00 0.42 0.35 o0.01 1.90
OP240T4 GMD 29.00 0.71 0.57 o0.01 2.50

Overall 349.00 0.45 0.39 o0.01 1.93

Table 5
Total spacing values for the core logging data

Borehole no. Rock type Mapped length (m) Total Joint Spacing (m)

Mean SD Min Max

HVGD002 GMD 169.00 0.52 0.52 o0.01 2.47

HVGD003 GMD/PB 169.00 0.40 0.41 o0.01 2.54
GTD55 GMD/PB 104.00 0.40 0.43 o0.01 2.95
ORGD001 GMD/PB 108.00 0.42 0.46 o0.01 2.58
HMGD009 GMD 76.00 0.51 0.44 o0.01 2.03
HMGD010 GMD 157.00 0.46 0.48 o0.01 3.48
HMGD011 GMD 21.00 0.39 0.37 o0.01 1.58
PAGD010 GMD/PB 228.00 0.50 0.42 o0.01 3.52
Combined data GMD and PB 1032.00 0.45 0.45 o0.01 3.52
596 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602

Table 6
Set spacing and normal set spacing values for the four main discontinuity sets

Set 1 Set 2 Set 3 Set 4

Set spacing (m) Mean 0.90 1.13 F F

(From core logging) SD 1.22 1.60 F F
Min o0.01 o0.01 F F
Max 11.30 8.91 F F

Set spacing (m) Mean 1.59 3.23 2.67 1.89

(From scanline mapping) SD 1.85 3.47 2.58 2.69
Min o0.01 0.10 o0.01 o0.01
Max 12.60 16.90 10.80 14.20

Normal set spacing (m) Mean 0.72 1.00 F F

(From core logging) SD 0.95 1.42 F F
Min o0.01 o0.01 F F
Max 7.99 7.83 F F

Normal set spacing (m) Mean F F 0.89 1.64

(From scanline mapping) SD F F 0.86 2.58
Min F F o0.01 o0.01
Max F F 3.60 14.16

400 100
90
350
80
300 70
Frequency

250 60
Frequency

50
200
40
150 30

100 20
10
50
0
0.1
0.3
0.5
0.7
0.9
1.1
1.3
1.5
1.7
1.9
2.1
2.3
2.5
2.7
2.9
3.1
3.3
0
0.1

0.4

0.7

1.3

1.6

1.9

2.2

2.5

2.8

3.1

3.4
1

Spacing (m)
Spacing (m) Fig. 4. Frequency histogram for total discontinuity spacing from
Fig. 3. Frequency histogram for total discontinuity spacing from core scanline mapping data.
logging data.

80.00
while the average SD is B20% greater for the core
logging data. This may be explained by the recording of 70.00

the more significant quartz veins as discontinuities 60.00

during the scanline mapping process, with these being
50.00
Frequency

subsequently included in the calculations for the average

total discontinuity spacing. Because it is unlikely that all 40.00
of the measured veins would cause breakage of diamond 30.00
core, the influence of the veins will have been more
20.00
significant in decreasing the spacing values determined
from scanline mapping. 10.00
The diamond holes were drilled across the predomi- 0.00
nant orientations of the major faults and shear zones.
0.10

0.70

1.30

1.90

2.50

3.10

3.70

4.30

4.90
5.50

6.10

6.70

7.30

7.90

The scanline mapping, however, was deliberately under-

Spacing (m)
taken along clean batter faces that were not significantly
affected by faulting and shearing. Therefore, while the Fig. 5. Frequency histogram for joint set 1 normal spacing.
 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602 597

Fig. 6. Frequency histogram for joint set 2 normal spacing. Fig. 7. Frequency histogram for joint set 4 normal spacing.

scanline mapping was generally undertaken along measurements versus the corresponding percentage
exposures of good rock mass condition, the boreholes deviation for the individual boreholes and scanlines
intersected both good and poor quality rock masses. presented in Table 8. The figure also indicates that as the
This is likely to have contributed to the higher SD for number of spacing measurements increases, the im-
the borehole spacing data. proved reliability achieved by taking more measure-
Both frequency histograms for the total discontinuity ments decreases. For example, if 500 more
spacing values (Figs. 3 and 4) follow the general shape measurements had been taken from borehole
of the exponential distribution, although Fig. 4 is more HVGD002, the reliability of the mean spacing estimate
poorly defined due to a smaller number of samples. This for those data would have increased by 53% (assuming
is further reflected in the small difference between the the same SD). However, if 500 measurements were
average and SD values for each data set (Table 7). added to the combined core logging data for which
The frequency histograms for the normal set spacing significantly more spacing measurements were available,
values of joint sets 1, 2 and 4 (Figs. 5–7) also appear to the reliability of the mean spacing estimate for that data
follow the general shape of the exponential distribution would have increased by only 14%.
but not as clearly as the histograms for the total The percentage error recommended in practice lies in
discontinuity spacing data. This is manifested in the the range 5–7% [8]. The percentage deviation for each of
larger differences between the average and SD values for the boreholes and scanlines is significantly greater than
the normal set data (also shown in Table 7). this recommended range at the 99% confidence level.
However, the percentage deviation for the combined
3.6. Inaccuracy and imprecision of the mean discontinuity core data falls within the recommended range at the
spacing estimates 99% confidence level, indicating that the reliability of
the combined data is good and provides a quantitative
Eq. (3) has been used to analyse the precision of the indication of the benefits of combining information
mean spacing estimates from the core logging and the from separate data sets.
scanline mapping data. All of the boreholes used in this
study were drilled in approximately the same orienta-
tion, allowing Eq. (3) to also be applied to the combined 3.7. The effects of orientation bias
core logging data set. The results are presented in
Table 8 as a percentage of the mean spacing estimates. The influence of orientation bias is reflected in the set
An example will assist to explain the information spacing values for the four main joint sets (Table 6). The
presented in the table: for borehole HVGD002, effects are particularly obvious for joint sets 1 and 2,
there is 95% confidence that the mean set spacing with the borehole data producing significantly smaller
estimate of 0.52 m lies within 16.57% of the true set spacing estimates for each set when compared to the
population mean. scanline mapping data. The orientations of most of the
As expected, the percentage deviation decreases as the scanlines form very acute angles with sets 1 and 2 while
sample size increases, indicating that a greater number most of the boreholes are at relatively high angles to
of spacing measurements will provide a more reliable these sets. The discontinuities from each set were
mean set spacing estimate. This is reflected in Fig. 8, therefore encountered less regularly by the scanline
which presents a chart of the number of spacing mapping process.
598 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602

Table 7
Difference between average and standard deviation values for the normal and set spacing

Mean SD Percent difference between mean and SD

Total spacing (core logging data) 0.45 0.46 2.22

Total spacing (NS striking scanline data) 0.48 0.46 4.17
Normal set spacing for joint set 1 0.72 0.95 31.94
Normal set spacing for joint set 2 1.00 1.42 42.00
Normal set spacing for joint set 4 1.64 2.58 57.32

Table 8
Precision of mean set spacing estimates

Borehole/ Rock type Spacing (m) Precision (as a percentage) of the mean spacing estimate for the confidence intervals indicated
scanline no.
Mean SD n 50% 60% 70% 80% 90% 95% 99%

HVGD002 GMD 0.52 0.52 140 5.70 7.12 8.76 10.83 13.90 16.57 21.77
HVGD003 GMD 0.39 0.39 264 4.15 5.18 6.38 7.89 10.12 12.06 15.85
PB 0.56 0.64 16 19.29 24.06 29.60 36.63 47.00 56.00 73.60
GTD55 GMD 0.39 0.42 240 4.69 5.85 7.20 8.91 11.44 13.62 17.91
PB 0.46 0.46 40 10.67 13.31 16.38 20.27 26.01 30.99 40.73
ORGD001 GMD 0.43 0.47 219 4.99 6.22 7.65 9.47 12.15 14.48 19.03
PB 0.36 0.39 24 14.93 18.62 22.91 28.35 36.38 43.34 56.96
HMGD009 GMD 0.51 0.44 97 5.91 7.38 9.08 11.23 14.41 17.17 22.57
HMGD010 GMD 0.46 0.48 308 4.01 5.01 6.16 7.62 9.78 11.65 15.32
HMGD011 GMD 0.39 0.37 27 12.32 15.37 18.92 23.41 30.03 35.79 47.03
PAGD010 GMD 0.50 0.39 348 2.82 3.52 4.33 5.36 6.88 8.20 10.77
PB 0.56 0.67 28 15.26 19.04 23.42 28.99 37.19 44.32 58.24
ON130T1 PB 0.24 0.23 54 8.80 10.98 13.51 16.72 21.45 25.56 33.59
ON130T2 PB 0.53 0.44 72 6.60 8.24 10.14 12.54 16.09 19.18 25.20
OY140T1 GMD 0.56 0.58 73 8.18 10.21 12.56 15.54 19.94 23.76 31.23
OY140T2 GMD 0.76 0.59 66 6.45 8.05 9.90 12.25 15.72 18.73 24.62
OY150T1 GMD 0.36 0.34 129 5.61 7.00 8.61 10.66 13.68 16.30 21.42
OP230T1 GMD 0.35 0.33 72 7.50 9.36 11.51 14.25 18.28 21.78 28.62
OP240T1 GMD 0.27 0.19 82 5.25 6.54 8.05 9.96 12.78 15.23 20.02
OP240T2 GMD 0.31 0.27 83 6.45 8.05 9.90 12.26 15.73 18.74 24.63
OP240T3 GMD 0.42 0.35 88 6.00 7.48 9.20 11.39 14.61 17.41 22.88
OP240T4 GMD 0.71 0.57 41 8.46 10.56 12.99 16.07 20.62 24.57 32.30
Combined core data PB & GMD 0.45 0.45 1751 1.61 2.01 2.48 3.06 3.93 4.68 6.16

4. Discontinuity frequency and rock quality designation

4.1. Background
80.00
Discontinuity frequency is the inverse of discontinuity
% deviation from population mean

70.00 spacing and will therefore also influence the block size
60.00 within a rock mass. Frequency can be defined in terms
50% confidence of occurrence per unit volume, unit area or unit length.
50.00 60% confidence
70% confidence Linear frequency is the simplest and most commonly
40.00 80% confidence
used method [3] and is generally defined as the number
90% confidence
30.00 95% confidence of discontinuities intersecting a unit length of a sampling
99% confidence
20.00
line, such as a scanline or drill core.
Hudson and Priest [11] reported that discontinuities
10.00
are never similarly distributed in all directions and, as a
0.00 consequence, frequency values will depend on the
0 100 200 300 400 direction of the mapping line. They provide formulae
Number of measurements
that allow the estimation of discontinuity frequency
Fig. 8. Precision of mean set spacing estimates. along any orientation for a given discontinuity set of
 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602 599

known orientation and spacing. The formulae allow found that RQD* can be calculated to within 6% of the
detailed analysis of frequency variation and can be used measured RQD using the proposed equation.
to estimate the directions and magnitudes of the
maximum and minimum frequency values for a given
rock mass. 4.2. RQD for the Oroya cutback analysis
Rock quality designation (RQD) is defined as the
percentage of diamond core recovered in intact pieces of RQD values measured from diamond drill core are
100 mm or more in length and is fundamental to the presented in Table 9. The measurements were split into
various rock mass classification schemes. Where dis- 1 m increments to allow an unbiased calculation of the
continuity spacing follows an exponential distribution, SD.
Priest and Hudson [4] proposed the following relation- The total discontinuity spacing values from the
ship between a theoretical RQD* and the mean north–south striking scanlines appear to follow an
discontinuity frequency per metre (l): exponential distribution, as discussed above. Conse-
quently, Eq. (5) has been applied to the scanline
RQD ¼ 100 e0:1l ð0:1l þ 1Þ: ð5Þ mapping data to produce the RQD* estimates presented
in Table 10. The number of discontinuities per metre (l)
Priest and Hudson [4] also proposed that the use of
and therefore the RQD* values were initially calculated
this equation in conjunction with scanline surveys on
for 1 m increments. The linear approximation provided
freshly exposed faces is more appropriate than the
by Eq. (6) was not applied to the scanline mapping data
determination of RQD from diamond core. This is
because the average number of discontinuities per metre
because core logging may tend to underestimate RQD
does not fall within the specified range.
values due to the formation of induced fractures during
the drilling process.
A plot of l versus RQD* is shown in Fig. 9 [4]. A 4.3. Discussion
good linear approximation to Eq. (5) within the range
6olo16 is provided by the tangent at the inflection The total average RQD* value is B4.5% higher than
point of the curve. The equation of the tangent is given the total average RQD value (see Tables 9 and 10). This
by may be explained by the influence of the foliation. Well-
RQD ¼ 110:4  3:68l: ð6Þ developed foliation planes may have broken the core but
may not have appeared as prominent discontinuities in
For meta-volcanic and meta-clastic rocks in Manito- batter faces. These planes may therefore have been
ba, Wallis and King [6] determined RQD from included in the determination of RQD from core logging
individual scanlines and drill core measurements and but may not have been mapped as discontinuities during
corresponding values of RQD* using Eq. (5). They scanline mapping.

Fig. 9. Relation between RQD and mean discontinuity frequency [4].

600 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602

Table 9
RQD estimates from the core logging data

Borehole no. Rock type Logged length (m) RQD

Mean Mode Median SD Min Max

HVGD002 GMD 169.00 95.00 97 95 4.08 60.00 100.00

HVGD003 GMD 150.00 79.00 100 95 15.53 0.00 100.00
PB 19.00 92.00 97 97 10.64 69.00 100.00
GTD55 GMD 74.00 92.00 100 97 14.24 32.00 100.00
PB 30.00 95.00 100 94 3.93 89.00 100.00
ORGD001 GMD 96.00 92.00 100 97 13.92 0.00 100.00
PB 12.00 87.00 97 94 12.93 50.00 98.00
HMGD009 GMD 76.00 96.00 100 98 5.90 83.00 100.00
HMGD010 GMD 157.00 97.00 100 100 8.03 30.00 100.00
HMGD011 GMD 21.00 95.00 100 94 6.52 72.00 100.00
PAGD010 GMD 186.00 93.00 100 97 7.21 42.00 100.00
PB 32.00 95.00 100 96 12.34 30.00 100.00

Overall 92.33 99.25 96.17 9.61 47.75 99.83

Table 10
RQD* estimates from the scanline mapping data

Scanline no. Rock type Mapped length (m) RQD*

Mean Mode Median SD Min Max

ON130T1 PB 13.00 93.00 94 94 5.41 81.00 100.00

ON130T2 PB 38.00 98.00 100 98 2.14 91.00 100.00
OY140T1 GMD 41.00 98.00 98 98 1.78 94.00 100.00
OY140T2 GMD 50.00 99.00 100 100 1.53 94.00 100.00
OY150T1 GMD 47.00 96.00 96 96 3.28 84.00 100.00
OP230T1 GMD 46.00 96.00 98 96 3.21 88.00 100.00
OP240T1 GMD 22.00 95.00 91 95 2.92 91.00 100.00
OP240T2 GMD 26.00 95.00 96 96 4.50 81.00 100.00
OP240T3 GMD 37.00 97.00 98 98 2.30 81.00 100.00
OP240T4 GMD 29.00 99.00 100 100 1.47 94.00 100.00

Overall 96.60 97.10 97.10 2.85 87.90 100.00

45 The effects of orientation bias will have caused the

40 diamond holes to intersect the predominant geological
35 orientations more regularly than the scanlines. Scanline
30
ON130T1 was mapped across the strike of the
Frequency

25
predominant structural orientations resulting in the
20
relatively low average RQD* value of 93%. The value
15
10
is close to the overall average value for the boreholes
5
and is significantly less than the average values obtained
0 from the other scanline surveys. However, scanline
60 65 70 75 80 85 90 95 100 OP230T1 was also mapped across strike, and its RQD*
RQD (%) value of 96% is not appreciably lower than those values
Fig. 10. Frequency histogram for RQD values logged from diamond obtained when mapping along strike. This may be due
hole HVGD002. to joint set 1 being relatively less developed in this area.
Prominent quartz veins were recorded as disconti-
nuities during the scanline mapping process and these
All of the diamond holes were drilled across the strike were subsequently included in the calculations of
of the predominant structural orientations, whereas RQD*. It is unlikely that all of the measured veins
eight out of the 10 scanlines were mapped approxi- would cause breakage of diamond core and therefore
mately parallel to the predominant geological structures. the influence of these features will have been greater in
 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602 601

the calculation of RQD* than the calculation of RQD. inexpensive when compared to diamond drilling.
If this is true, a more realistic approach to the treatment Therefore attempts should be made in further work
of quartz veins would lead to even higher values of to enhance the accuracy of the method through an
RQD* due to a reduction in the measured number of improved understanding of the influence of quartz
discontinuities per metre. veins and foliation on the rock mass.
The SD for the RQD values is significantly higher for * The percentage deviation for the combined core
the core logging data (Tables 9 and 10). This is most logging discontinuity spacing data falls within the
likely due to the diamond holes intersecting major faults recommended error percentage range at the 99%
and shear zones while the scanline mapping was confidence level. The reliability of the combined core
undertaken along exposures of competent rock. This logging data is therefore adequate for use in the
effect is also reflected in the lower minimum RQD engineering analysis of the cutback design.
values for the core logging data.
The information in Tables 9 and 10 shows that the For the purposes of subsequent engineering analysis
relative positions of the mean, mode and median values in the cutback, only the RQD values determined from
for the RQD and RQD* are as expected for a left- core logging were used in the estimation of the rock
skewed distribution, typical of a relatively high quality mass shear strength for the Oroya cutback design. The
rock mass. This is clearly illustrated in Fig. 10, which RQD* values determined from scanline mapping were
presents by way of example the RQD distribution for not used for the following reasons:
borehole HVGD002.
* The standard method for determining RQD is the use
of diamond core and a total of 1022 m of core were
available for this investigation. It was therefore
5. Conclusions
considered unnecessary to complement these data
using other (indirect) techniques.
The following conclusions of general interest have * The complications associated with the quartz veins
been made:
and the foliation planes both relate to the difficulty in
* The total discontinuity spacing values for the study determining which features should be included as
area clearly follow the shape of the exponential discontinuities during scanline mapping. This diffi-
distribution, which is consistent with many previous culty is eliminated when logging core by recording
workers. only those natural features that cause breakage of the
* Although not as clearly defined, the normal set core.
spacing values also appear to follow the exponential * The effect of the foliation planes in decreasing the
distribution. RQD values (as determined from core logging) may
* The total average calculation for RQD* in this study provide a more realistic indication of the rock mass
is within 5% of the total average estimate for RQD. quality since foliation will be encountered throughout
Consequently, the method proposed by Priest and much of the cutback area.
Hudson [4] provides a valuable tool that would be * The intersection of the diamond holes with major
particularly beneficial for geotechnical investigations faults and shear zones also provides a realistic
where diamond core is not readily available. indication of the rock mass conditions.
* On average, the RQD values determined from core
Conclusions specific to the study of the Oroya cutback logging are lower than those determined from scan-
design are as follows: line mapping. The core logging data therefore
* Complications in the discontinuity spacing results are provides a more conservative approach.
created by the influence of quartz veining and
foliation planes. The veins appear to have influenced However, the potential exists to increase the reliability
the scanline mapping data to a greater extent than the of the RQD* estimates in the study area even further
core logging data, while the foliation appears to have through an improved understanding of the influence of
influenced the core logging data to a greater degree. quartz veining and foliation on the rock mass.
* Because of these complications, only those disconti-
nuity spacing values that were obtained from core
logging were used in the geotechnical engineering Acknowledgements
analysis of the Oroya cutback.
* With more than 1780 discontinuity spacing measure- This paper is based on part of a research thesis [12]
ments being available from the core logging data, the submitted to the Western Australian School of Mines,
exclusion of the scanline mapping data is acceptable. Curtin University of Technology. The authors wish to
However, the scanline mapping technique is relatively acknowledge the support of KCGM (who employed the
602 D.R. Wines, P.A. Lilly / International Journal of Rock Mechanics & Mining Sciences 39 (2002) 589–602

first author during the period that this work was [6] Wallis PF, King MS. Discontinuity spacings in crystalline rock.
undertaken) and for that company’s permission to Int J Rock Mech Min Sci & Geomech Abstr 1980;17:63–6.
publish this paper. [7] Hadjigeorgiou J, Lessard JF, Villaescusa E, Germain P. An
appraisal of structural mapping techniques. In: Rossmanith HP,
editor. Mechanics of jointed and faulted rock. Rotterdam:
Balkema, 1995. p. 191–7.
References [8] Sen Z, Kazi A. Discontinuity spacing and RQD estimates from
finite length scanlines. Int J Rock Mech Min Sci & Geomech
[1] Clout JMF, Cleghorn JH, Eaton PC. Geology of the Kalgoorlie Abstr 1984;21:203–12.
Gold Field. In: Hughes FE, editor. Geology of the mineral [9] Hudson JA, Priest SD. Discontinuities and rock mass geometry.
deposits of Australia and Papua New Guinea, The Australasian Int J Rock Mech Min Sci & Geomech Abstr 1979;16:339–62.
Institute of Mining and Metallurgy, Melbourne, 1990. p. 411–31. [10] Priest SD, Hudson JA. Estimation of discontinutiy spacing and
[2] Giani GP. Rock slope stability analysis. Rotterdam: A.A trace length using scanline surveys. Int J Rock Mech Min Sci &
Balkema Publishers, 1992. Geomech Abstr 1981;18:183–97.
[3] Priest SD. Discontinuity analysis for rock engineering. London: [11] Hudson JA, Priest SD. Discontinuity frequency in rock masses.
Chapman & Hall, 1993. Int J Rock Mech Min Sci & Geomech Abstr 1983;20:73–89.
[4] Priest SD, Hudson JA. Discontinuity spacings in rock. Int J Rock [12] Wines DR. Aspects of rock slope engineering at the Fimiston
Mech Min Sci & Geomech Abstr 1976;13:135–48. Open Pits. Unpublished MEng thesis, Mining Engineering
[5] Call RD, Savely JP, Nicholas DE. Estimation of joint set Department, Western Australian School of Mines, Curtin
characteristics from surface mapping data. In: Hustrulid WA, University of Technology, 2000, 333pp.
editor. Monograph on rock mechanics applied in mining.
New York: AIME, 1976. p. 65–73.

View publication stats